soliwilos_ has quit [Remote host closed the connection]
soliwilos has joined #kisslinux
midfavila has quit [Remote host closed the connection]
<testuser[m]12>
Hi
<wael[m]>
\o
<Torr>
Hey
<virutalmachineus>
HI sir
<Torr>
Leaving here
<Torr>
See ya
Torr has quit [Quit: leaving]
<wael[m]>
cya
<virutalmachineus>
see ya brother, may kiss be with you
ioraff has quit [Ping timeout: 268 seconds]
ella-0 has joined #kisslinux
ella-0_ has quit [Ping timeout: 256 seconds]
ejjdhfjsu has joined #kisslinux
ioraff has joined #kisslinux
rohan has quit [Ping timeout: 256 seconds]
rohan has joined #kisslinux
rohan has quit [Ping timeout: 256 seconds]
ejjdhfjsu has quit [Remote host closed the connection]
fitrh has joined #kisslinux
ioraff has quit [Quit: ioraff]
fitrh_ has joined #kisslinux
fitrh has quit [Ping timeout: 256 seconds]
sad_plan has joined #kisslinux
<sad_plan>
hi
<sad_plan>
wael[m]: toybox is actually *less* than half of busybox's size
<sad_plan>
busybox is 205k, where as toybox is 72k
sad_plan has quit [Quit: nyaa~]
sad_plan has joined #kisslinux
fitrh_ has quit [Read error: Connection reset by peer]
fitrh has joined #kisslinux
<wael[m]>
damnnn
<wael[m]>
wait how tf is it 205k
<wael[m]>
for me its 1 meg and toybox is 460k
<wael[m]>
thats probably because im on glibc
<sad_plan>
perhaps. things tend to get way larger on glibc.
<sad_plan>
running cloc on both of them will give us those numbers
<sad_plan>
the size will also differ from the config you use
<wael[m]>
i used the default configuration on both
<sad_plan>
ok, never used the def config on toybox, but the current config is a little less than 1m iirc. 800k or something
<sad_plan>
or 628k to be correct actually. so not that bad. I still have some tinkering to do with it to make it work correctly though, or atleast sufficently in any case
<wael[m]>
so true
<sad_plan>
yeah. did you build toybox's init?
<wael[m]>
no i just want its coreutilities
<sad_plan>
ah ok
<wael[m]>
is it similar to busybox?
<wael[m]>
if its possible i would prefer to just swap out everything to toybox
<sad_plan>
more or less. I just havent been able to get it to boot with the inittab for some reason. Im not sure why. I skipped it earlier and just used shinit instead
<wael[m]>
shinit? you can have a shell as an init?
<sad_plan>
should be possible. look at dylins repo for it. swapon/off doesnt have the -a flag, so youd have to remove it, or specify what swap partition its on
<sad_plan>
yep
<wael[m]>
i mean, how would services get ran then
<sad_plan>
it just runs the init, nothing else
<sad_plan>
youd have to manually start your serivice manager using a bootscript
<wael[m]>
ohh yeah that makes sense, that being toybox/busybox runit yes?
<sad_plan>
the patch might needs some adjustment, so just apply it manually. its not a large patch anyway
<sad_plan>
testuser[m]12: why?
<wael[m]>
testuser: are you going to drop libxslt? its kinda outdated and the old version does not exist anymore ( upstream, not repo )
<testuser[m]12>
its orphaned
<wael[m]>
sad
<testuser[m]12>
sad_plan no ublock
<wael[m]>
there is lariza which has adblocking i believe
<sad_plan>
is there a better way to build gcc when gcc's bins are static? Ive set up so it builds the bins statically, but including shared libs. every time I build a different version, I always get errors with the shared libs. to build the recent one, I have to revert to a build from a tarball, then rebuild again
<sad_plan>
testuser[m]12: /etc/hosts? :p ublock is great, not gonna argue on that one though.
<testuser[m]12>
and no extension or built in feature like chromium for dark mode
<wael[m]>
tbh i just dont want firefox consuming all my damn memory
<wael[m]>
if only it was low like webkit2gtk i'd be happy
<testuser[m]12>
If anyone actually uses them they'll know when kiss tells them it doesn't exist in Any repo
<sad_plan>
darkmode can be enable in one of the later commits. or just created manually. im not saying surf is a better browser, by any stretch. I suppose its a more principal thing or something.
<wael[m]>
sad_plan: does ruby build for you on libressl?
<sad_plan>
wael[m]: not on my laptop, but on my desktop it does. for some reason. try an older version perhaps
<sad_plan>
testuser[m]12: I can take some of those that I use, but i.e go which I previously commited to, is way above my head, if theres something in particular which must be done. bumping version numbers is easy I mean :p
<wael[m]>
sad_plan: ive tried all ruby versions till 3.0 and it cant build, why is it a dependency of webkit2gtk anyway?
<sad_plan>
webkit2gtk has alot of ruby in it appearantly
<wael[m]>
damn wtf what for
<testuser[m]12>
Codegen
<virutalmachineus>
webkit2gtk is bloat
<sad_plan>
build it in a kiss chroot, and just install the tarball afterwards from your user. extract it in /tmp i.e, and just build ruby in there.
<testuser[m]12>
Maybe
<wael[m]>
sad_plan: why a kiss chroot? i believe ruby provides binaries so i could use those
<virutalmachineus>
use vms free from the bloat
<wael[m]>
vms != a browser
<virutalmachineus>
browsers are huge
<sad_plan>
because that way you can actually build it. I was going to say those bins arent made for musl, but I came to think of that you use glibc anyway
<virutalmachineus>
might as well run them in vms
<wael[m]>
:/
<sad_plan>
in any case, building ruby doesnt take that long. just extract a kiss tarball somewhere in a folder. build ruby, run kiss i $(path to ruby tarball)
<sad_plan>
has to be ran as root though
<testuser[m]12>
lol
<wael[m]>
this is painful
<testuser[m]12>
vms are adding more bloat if anything
<sad_plan>
it is, but its what Ive got currently as a workaround :p
<wael[m]>
is libxcrypt on kiss
<wael[m]>
aka libcrypt.so.2
<virutalmachineus>
qemu is small
<virutalmachineus>
even so if you use you're own .mak
<virutalmachineus>
* even more so if you use your own .mak
fitrh has quit [Quit: fitrh]
fitrh has joined #kisslinux
fitrh has quit [Client Quit]
<sad_plan>
testuser[m]12: did you read my comment on gcc?
<sad_plan>
yeah. gcc will bootstrap if the versions differ too much
<testuser[m]12>
u can disable it
<sad_plan>
Im assuming it does, as I was on 11.2 and now building 12.2
<testuser[m]12>
wael[m]: Build tools necessary to build the compiler.
<testuser[m]12>
Perform a comparison test of the stage2 and stage3 compilers.
<testuser[m]12>
Perform a 3-stage bootstrap of the compiler. This includes building three times the target tools for use by the compiler such as binutils (bfd, binutils, gas, gprof, ld, and opcodes) if they have been individually linked or moved into the top level GCC source tree before configuring.
<testuser[m]12>
Build runtime libraries using the stage3 compiler from the previous step.
<sad_plan>
by this isnt gcc more or less able to build itself? I seem to recall reading that someplace..
<sad_plan>
nvm
<sad_plan>
what do you guys think of the framework laptop?
<wael[m]>
i think its quite burger
<sad_plan>
what
<wael[m]>
sorry i meant its quite nice
<sad_plan>
ah. yes, I also think its really nice. Im really considering switching out my laptop for one
<wael[m]>
can it be librebooted?
<sad_plan>
no, but theyre talking about coreboot though. which is nice
<sad_plan>
its not announced yet though, but its roumored in any case
<wael[m]>
what do you think about the pinebook then?
<sad_plan>
its also talk of other achetectures aswell. ARM, risc, amd etc
<sad_plan>
its a neet project, but massivly underpowered. ive been thinking about the pinephone for a while, but I dunno.
<sad_plan>
afaik, all of their devices are really slow
<sad_plan>
which is fine I suppose, as theyre mostly for tinkering anyway
<wael[m]>
personally ive seen the 'linux' phones are extremely overpriced and id rather go with a lineage/crdroid microg-ed phone
<sad_plan>
however, a linux phone is somewhat appealing
<sad_plan>
they are. atleast the other one. I cant think of the name atm, but you probably know which one. the not-pinephone
<wael[m]>
also the price to performance is extremely garbage
<sad_plan>
purism
<sad_plan>
indeed ^
<sad_plan>
the librem 5. was the phone I thought about. which is even worse with price to performance ratio...
<sad_plan>
im sure the fxtec pro1-x is a better option tbh
<sad_plan>
or a fairphone
<sad_plan>
or if you get a oneplus 6 to install a distro
<wael[m]>
there is postmarket os
<sad_plan>
yeah, postmarked is really nice imo. its like alpine, but on a phone. which is aaalmost kiss
<wael[m]>
kiss linux on a phone
<sad_plan>
you can use sxmo on pmos and youll get a suckless phone more or less. in any case a suckless IU
<sad_plan>
s/IU/UI/
<wael[m]>
also i heard the battery is absolutely shit
<sad_plan>
yeah, probably
<sad_plan>
I dunno on all the ported builds to various phones, but the pinephone and such is absolutly garbage
<sad_plan>
wtf. gcc error out on there not being a /gcc-build/c++tools dir.
<wael[m]>
i believe dilyn has made a kiss-community custom git repo before
<wael[m]>
but moving to codeberg is going to be an interesting choice because everyone is on github nowadays
<sad_plan>
yeah, its one his linode server. but only he has access iirc
<sad_plan>
wael[m]: itll be a mirror iirc. so people can still use github. or just email, and send patches on the mailing list or w/e
<wael[m]>
so whats the point lol
<illiliti>
testuser[m]12: nice
<illiliti>
i'm slowly moving to codeberg too
<illiliti>
my new projects will be there
<illiliti>
however, i'll keep maintaining current projects on github cuz moving them to codeberg is too much headache for me and package maintainers
<sad_plan>
isnt it just as easy as creating a mirror? and pushing it to the new mirror?
<illiliti>
i don't want to maintain a mirror
<illiliti>
either github or codeberg. not both
<sad_plan>
cant you just mirror it first, then once you pushed everything to codeberg, just delete it on github?
<illiliti>
that would be a disaster for package maintainers
<illiliti>
cuz they would need to update sources and stuff. i don't want to put this needless burden on them
<sad_plan>
touche. I didnt think of people actually using your stuff.
<illiliti>
well, they are actually
<sad_plan>
I knoow. I just forgot for a minute :c
fitrh has joined #kisslinux
<illiliti>
one option can be if first mirror it, then make an announcement about mirror, then submit patches to every distro that maintain my shit and archive github repo
<illiliti>
this way nobody will suffer but me
<illiliti>
well almost, it's still substantial burden
<sad_plan>
but youll end up with all of your shit on codeberg that way. instead of having some stuff on gh and some other on codeberg
<sad_plan>
which is more or less what you want
<illiliti>
yep
sad_plan has quit [Quit: nyaa~]
fitrh has quit [Read error: Connection reset by peer]
fitrh has joined #kisslinux
dilyn has joined #kisslinux
<dilyn>
illiliti: just setup a git hook which pushes to the github mirror whenever you push to codeberg?
<testuser[m]12>
and disable issues and pr on shithub
<illiliti>
again, i don't want to maintain a mirror in any way
<illiliti>
i want to trash away github altogether
<illiliti>
wipe it, nuke it
<illiliti>
github doesn't support ipv6 for pushing. i won't be able to maintain a mirror in ipv6-only environment
<illiliti>
it's one of the reasons why mirror is no go
<dilyn>
oho i see i see
<illiliti>
yeah
<illiliti>
mirror on github implies that i would have to fight with github dumbness
<testuser[m]12>
ig they want to support some system from the 1800s
<illiliti>
fix: drop this garbage and use mkdir -p
<testuser[m]12>
how would u even use any script on such a system lol
<illiliti>
good question
<dilyn>
illiliti: creating a pc file for llvm is basically trivial, given that all the information is known at build-time (or queryable at run-time with llvm-config)
<dilyn>
so there's nothing really stopping us there
<illiliti>
yeah, i just haven't done that yet
<testuser[m]12>
can you portably sed within a sed?
<illiliti>
afaik it's possible to invoke commands from sed expession, so i think ye
<illiliti>
good start, but yeah general approach is to generate them
<illiliti>
or just use posix make
<illiliti>
or muon if you need more complex logic
<illiliti>
everything else is crap
<phoebos>
posix make improves my mental health
<phoebos>
unlike every other build system
<testuser[m]12>
What's bad about this syntax? It's very declarative
<testuser[m]12>
95% of the time you just need to create targets and set properties on them.
<phoebos>
ninja?
<phoebos>
yeah ninja is nice, but almost always comes with a generator
sad_plan has joined #kisslinux
<phoebos>
specifically, posix make makes me happy to use more than write
<sad_plan>
Great start indeed. Ill look mord into that file. I had trouble including a header file. I just tried to build scroll, as its a simple 2 file program. But yeah..
<testuser[m]12>
Know that some people, somewhere, are having the time of their lifes, rolling around in stacks of money earned through billed CMake consultancy hours, books, and conference talks.
<phoebos>
lol
sad_plan has quit [Quit: Quit]
soliwilos has quit [Remote host closed the connection]
soliwilos has joined #kisslinux
soliwilos has quit [Remote host closed the connection]
soliwilos has joined #kisslinux
ioraff has quit [Remote host closed the connection]
<midfavila>
on the topic of kiss, does anyone here run a compiler other than gcc as their primary one?
<midfavila>
as I pare my system down I'm considering attempting to replace the GNU collection and binutils
<midfavila>
i ran a cloc on binutils today and the tarball that I grabbed had over 4mil lines of code xwx
ioraff has joined #kisslinux
<Torr>
There's a fork which uses solely Llvm iirc.
<midfavila>
(i forgot that llvm exists)
<midfavila>
i suppose i should have been a little more precise -- something like pcc, tcc or cproc is more what I'm interested
<Torr>
It's rather bloated too, so not a big improvement over Gcc.
<midfavila>
s/interested/interested in/
<Torr>
Not that I know of.
<midfavila>
rip
<Torr>
That's your chance midfavila.
<midfavila>
yeah any day now :v
<ioraff>
not really an improvement at all. compiles faster, but binaries are slower. OSes picked it up for licensing reasons.
<Torr>
Pretty much
<midfavila>
eh, i'm not super concerned about raw performance, tbqh. it's more the mechanical complexity of the system as a whole that bothers me
<midfavila>
...oh, on the topic of massive development toolchains and accoutrement, you guys might find mold interesting
<midfavila>
"modern ld"
<midfavila>
apparently it's like 8x faster than llvm's ld
<midfavila>
super parallelized or something?
<ioraff>
that's fair. but stripping out all of the compiler flags that only gcc and clang support requires one to completely rewrite builds, like oasis
<ioraff>
mold is in extra/
<midfavila>
oh, guess i'm late to the party then
<midfavila>
i haven't tracked upstream kiss for a while
<midfavila>
i'm playing in the corner by myself ;w;
<ioraff>
what do you use?
<midfavila>
i forked kiss around a year ago
<midfavila>
so i'm just kind of limping along on that
<midfavila>
it mostly works, even
<midfavila>
still using libressl and i've replaced busybox with suckless and bsd tools, and swapped the zlib implementation, among other things